Robotics Deployment / Sim2Real Engineer

Responsibilities

• Deploy, debug, and validate robotics models on physical platforms, including robotic arms, grippers, cameras, sensors, controllers, and other hardware systems.
• Build and maintain real-robot data-collection workflows, including teleoperation, task-scene setup, trajectory recording, synchronisation of image, state, and action data, data cleaning, and format conversion.
• Deploy VLA, WAM, and other robot-policy models for inference on real robotic arms, and conduct closed-loop testing.
• Contribute to the Sim2Real pipeline, including simulation-environment setup, task modelling, sensor simulation, dynamics-parameter tuning, policy transfer, and real-robot validation.
• Debug robotic-arm platforms, including integration of ROS 2, MoveIt, control interfaces, and safety mechanisms for systems such as Franka, Agibot, and Piper.
• Configure and debug robotic-vision systems, including camera calibration, hand-eye calibration, multi-camera synchronisation, and workspace calibration.
• Resolve engineering issues encountered in real-robot deployment, including control latency, action limiting, safety protection, cable management, end-effector debugging, and sensor stability.

Qualifications

• Bachelor’s degree or above in robotics, automation, computer science, electronic engineering, mechanical engineering, or a related field.
• Familiarity with Linux development environments, proficiency in Python, and working knowledge of C++.
• Familiarity with ROS or ROS 2, and understanding of common robotics toolchains such as TF, URDF, MoveIt, ros_control, and ros2_control.
• Hands-on experience debugging physical robot hardware, with the ability to independently integrate and troubleshoot robotic arms, grippers, cameras, sensors, and related equipment.
• Understanding of fundamental concepts in robotic-arm motion control, end-effector pose control, trajectory planning, and joint-space and Cartesian-space control.
• Familiarity with at least one simulation platform, such as Isaac Sim, MuJoCo, Gazebo, PyBullet, or Isaac Gym.
• Experience with cameras and calibration, such as RealSense, ZED, industrial cameras, AprilTag, ChArUco, and hand-eye calibration.
• Strong engineering and deployment capability, with the ability to handle instability in real environments, including occlusion, calibration errors, communication failures, hardware limits, and safety issues.
• Ability to independently drive an end-to-end task, from scene setup and data collection to model deployment and real-robot testing.

Preferred Qualifications

• Hands-on experience with physical platforms such as Franka, Agibot, or Piper.
• Experience deploying or reproducing VLA models such as OpenVLA, RT-1/RT-2, ACT, Diffusion Policy, RDT, or GR00T, or WAM models such as FastWAM.
• Experience with imitation learning, reinforcement learning, robotics-dataset collection, or teleoperation-system development.
• Experience building and debugging laboratory robotics platforms.

Role Objective

After joining, the successful candidate will help build real-robot data-collection and model-deployment platforms, completing the full workflow from task definition, scene setup, and teleoperated data collection to data processing, model deployment, and real-robot inference validation. We are looking for someone who can not only write code, but also bring robotics systems online, make them run, and keep them stable.